Palmprint Feature Extraction Method Based on Rotation-invariance

摘要

Feature extraction is one of most basic problems in the research of palmprint recognition. Extracting effective palmprint feature is the crucial problem in the field of palmprint recognition. There was a research focus on how to select the feature. The lacking of main orientation for palmprint recognition system will lead to an incorrect feature extraction and matching. In order to extract more precise palmprint feature, a new method for feature extraction of palmprint was proposed in the research of palmprint recognition, which could improve the efficiency of identification. Firstly, calculating the main orientation of the whole image, and then adjusting the gradient of each pixel according to the main orientation to ensure this method has rotation invariance. Secondly, combining with the method of histogram of oriented gradient and dominant orientation. Finally, the feature value of palmprint was obtained. A reasonable threshold is set to estimate the similarity between the experimental images and the sample images. The experimental results showed that the method proposed in this paper can improve the efficiency of identification.